 |
 |
On Dec 17, 2004, at 7:11 AM, Fabrizio Furano wrote: > > rereading this message, I realized that this scenario is very similar > to the tricky one that Andy and I debugged and fixed together last > week. So, it might be useful for us to know: > - which client/server verison you are using (or the head of which day > if you are used to take the cvs head) I am using the release 20041214. I don't know the internal details of your CVS repository. > - what the client side is doing (xrdcp or some other prog?) In > particular which flags/options you specified in the Open request. The client side is a Suez program using the XrdClient interface. From what I can understand of Chris's code he just calls m_client->Open(O_RDONLY,0);. Thereafter the code does some Read() calls and a Close(). Additional Information: 1. This bug may not show up for several hours after starting xrootd, but sometimes within minutes. 2. After adding XrdClient debug level 1 on the client side, I can't reproduce the bug after 4 hours of trying. This suggests it may be timing-related. 3. I think the real problem must be in the xrootd server or olbd server, because once it gets into this mode, it stays in this mode until the server is restarted. Restarting the client doesn't fix it. That means that even if invalid behavior in the client triggers this behavior, the server is not rejecting the invalid behavior. I will keep trying to reproduce this with debug enabled. Should I move to the latest version that Peter released yesterday and see if it is fixed there? -- Gregory J. Sharp email: [log in to unmask] Wilson Synchrotron Laboratory url: http://www.lepp.cornell.edu/~gregor Dryden Rd ph: +1 607 255 4882 Ithaca, NY 14853 fax: +1 607 255 8062